A Fungal Symbiont of Plant-Roots Modulates Mycotoxin Gene Expression in the Pathogen Fusarium sambucinum

Fusarium trichothecenes are fungal toxins that cause disease on infected plants and, more importantly, health problems for humans and animals that consume infected fruits or vegetables. Unfortunately, there are few methods for controlling mycotoxin production by fungal pathogens. In this study, we isolated and characterized sixteen Fusarium strains from naturally infected potato plants in the field. Pathogenicity tests were carried out in the greenhouse to evaluate the virulence of the strains on potato plants as well as their trichothecene production capacity, and the most aggressive strain was selected for further studies. This strain, identified as F. sambucinum, was used to determine if trichothecene gene expression was affected by the symbiotic Arbuscular mycorrhizal fungus (AMF) Glomus irregulare. AMF form symbioses with plant roots, in particular by improving their mineral nutrient uptake and protecting plants against soil-borne pathogens. We found that that G. irregulare significantly inhibits F. sambucinum growth. We also found, using RT-PCR assays to assess the relative expression of trichothecene genes, that in the presence of the AMF G. irregulare, F. sambucinum genes TRI5 and TRI6 were up-regulated, while TRI4, TRI13 and TRI101 were down-regulated. We conclude that AMF can modulate mycotoxin gene expression by a plant fungal pathogen. This previously undescribed effect may be an important mechanism for biological control and has fascinating implications for advancing our knowledge of plant-microbe interactions and controlling plant pathogens.


Introduction
Secondary metabolites are compounds produced by many organisms including filamentous fungi. These compounds include pigments, toxins, plant growth regulators, antibiotics and numerous compounds used for pharmaceutical purposes [1]. Many fungal secondary metabolites increase the fitness of the organisms toward in adverse environmental conditions [2,3]. The fungal genus Fusarium (teleomorph: Giberrella) consists of over 70 species, many of which are plant pathogens, and some species can produce secondary metabolites, known as mycotoxins, that are toxic to humans and animals [1,4]. Trichothecenes are a family of terpenederived mycotoxins that impact human, animal [5] and plant [6].
Trichothecene biosynthetic gene clusters have been characterized in F. graminearum and F. sporotrichioides, [1,7]. In both species, there is a core cluster of 12 genes that are involved in the biosynthesis, regulation or transport of trichothecenes. These genes are: TRI5 (encoding a terpene synthase); TRI4, TRI11 and TRI13 (encoding cytochrome P450 monooxygenases); TRI3 and TRI7 (encoding acetyl transferases), TRI8 (encoding an esterase), TRI6 and TRI10 (proposed to be regulatory genes) and TRI12 (encoding a transporter). In addition, there are also two genes, TRI9 and TRI14 with unknown functions. The biosynthetic pathway for all trichothecenes begins with a cyclization of farnesyl pyrophosphate to produce the hydrocarbon trichodiene [8].
Control strategies against fungal pathogens that produce mycotoxins are mainly based on the use of fungicides [9], although biological-control agents [10] and plant-resistant varieties have been reported. However, the use of AMF to control mycotoxin-producing microorganisms has not been previously reported. AMF inhabit plant roots and not only form a symbiotic association with most plant species but also interact with a wide range of other soil organisms such as soil bacteria [11]. AMF are well known to promote plant growth and are largely used as commercial inoculants as bio-fertilizers worldwide. Many literature reports showed that AMF can reduce the incidence and severity of root diseases and protect plants against soil-borne pathogens [12]. However, the mechanisms by which AMF may act as biological control agents are not known. Three mechanisms have been hypothesized: soil microbial community changes, antagonisms and stimulation of plant defenses. AMF can inhibit or promote soil microorganisms [11]. It has been reported that soluble substances released by the extraradical mycelium of Glomus intraradices grown in vitro, stimulated both the growth of Pseudomonas chlororaphis and the germination of Trichoderma harizianum conidia.
In contrast, the germination of Fusarium oxysporum f. sp. chrysanthemi conidia was reduced in the presence of the AMF extract [13].
In this study, we identified 16 isolates belonging to the genus Fusarium from naturally infected potato plants and confirmed that nine produced trichothecenes. One isolate, F. sambucinum strain T5 induced a rapid wilting and yellowing that resulted in plant death and was selected for further studies. The objective of this work was to assess the effect of the AMF G. irregulare on growth and expression of TRI genes by F. sambucinum. We addressed three specific points: i) trichothecene production by F. sambucinum using molecular and chemical approaches; ii) effect of the AMF G. irregulare on the growth of trichothecene-producing F. sambucinum in vitro; and iii) G. irregulare-induced modulation of trichothecene gene expression in F. sambucinum.

Results and Discussion
Characterization of TRI genes of F. sambucinum We isolated sixteen strains of Fusarium from roots and tubers of naturally infected potato plants, and found that nine of these produced trichothecenes. We chose F. sambucinum strain T5 as a model for this study because it was the most aggressive strain when tested on potato plants, inducing a rapid wilting and yellowing that resulted in plant death ( Figure S2). This strain produced 4,15diacetoxyscirpenol (DAS) when grown in liquid culture ( Figure 1). We used ITS regions of ribosomal rRNA genes and morphology to confirm its identity. Nucleotide BLAST search showed 100% sequence identity with Gibberella pulicaris strain NBAIM: 455 (anamorph: F. sambucinum). This strain is a causal agent of dry rot of tuber crops [14]. We used degenerate primers (Table 1) and DNA from F. sambucinum to amplify fragments corresponding to five trichothecene genes (TRI5, TRI4, TRI101, TRI3, and TRI11) involved in production of DAS ( Figure 2). PCR gave bands of the expected sizes ranging from 1.1 to 1.5 Kb. These PCR products were sequenced and, when compared to the NCBI database using nucleotide BLAST searches, showed 98% and 97% sequence identity with trichothecene biosynthetic genes of F. sporotrichioides and Gibberella zeae, respectively.
To study the impact of G. irregulare on the growth of F. sambucinum, we used confrontation cultures using an in vitro system. We found that the growth of F. sambucinum was significantly reduced in the presence of G. irregulare isolate DAOM-197198 compared with controls that consisted of carrot roots without G. irregulare or F. sambucinum alone ( Table 2). G. irregulare significantly reduced the growth of F. sambucinum after 3, 5, 7 and 15 days ( Figure 3). Interestingly, inhibition of the growth F. sambucinum was associated with morphological changes in the presence of G. irregulare compared with controls ( Figure 3). These changes were observed when G. irregulare hyphae reached F. sambucinum mycelium, suggesting that G. irregular may produce compounds that interfere with normal the growth of F. sambucinum.
To test whether G. irregulare modulates the expression of TRI genes, we carried out real-time qRT-PCR assays on TRI4 and TRI5 on total RNA extracted from F. sambucinum grown alone or confronted with G. irregulare isolates DAOM-197198 and DAOM-234328 during 3 and 5 days. TRI4 and TRI5 genes encode P450 oxygenases and trichodiene, respectively [15]. We used ß-tubulin and the translation factor EF1a as reference genes for RT-PCR assays. Figure 4 shows relative expression patterns of TRI4 and TRI5 genes when F. sambucinum was confronted with the two isolates of G. irregulare. Interestingly, the relative expression of TRI5 in F. sambucinum was up-regulated (p,0.001) by a factor of 17 and RT-PCR doi:10.1371/journal.pone.0017990.t001 8 after 3 days of confrontation with G. irregulare isolates DAOM-197198 and DAOM-234328, respectively. TRI5 encodes a trichodiene synthase and is the first enzyme involved in the trichothecene biosynthesis pathway [1]. In contrast, the relative expression of TRI4 (encoding a multifunctional P450 oxygenase) was down-regulated (p,0.003) by a factor of 0.46 and 0.43 when confronted with isolates DAOM-197198 and DAOM-234328, respectively. Relative expression of TRI5 and TRI4 genes was not affected (p.0.22) when F. sambucinum was grown alone or confronted with carrot root without G. irregulare (Figures 4A; 4C). We used carrot root as a control because G. irregulare is an obligate biotroph that requires plant roots for its culture. These controls clearly show that over-expression of TRI5 and down-regulation of TRI4 was due to G. irregulare. Interestingly, the two isolates of G. irregulare showed different modulation levels on the expression of TRI5 and TRI4 genes in F. sambucinum. This difference in the response of G. irregulare isolates could be explained by their genetic composition. It is well documented in the literature that AMF have a high intra-and inter-isolate genetic diversity [16,17,18]. Recently, Angelard and coworkers [19] showed that segregated lines of the AMF G. irregulare (formerly G. intraradices) obtained from crossing two genetically different parents have a huge effect on host-plant response [19].
To test the effect of G. irregulare on other trichothecene biosynthetic genes, we carried out additional RT-PCR assays on TRI6, TRI13, and TRI101, using an experimental setup similar to that used for TRI4 and TRI5 (Fig. 4B and D). Relative expression of the transcription factor TRI6 gene was only up-regulated (p,0.03) by a factor of 2.8 and 3.6, after 3 and 5 days of confrontation with G. irregulare DOAM-234328, respectively ( Fig. 4B and D). Surprisingly, isolate DOAM-197198 produced no significant effect (p.0.08) on relative expression of TRI6 from F. sambucinum. This finding strongly supports genetic heterogeneity and phenotypic differences among isolates of G. irregulare. TRI6 encodes a zinc finger protein involved in regulation of trichothecene biosynthesis [20] with expression of TRI5 and TRI4 genes dramatically reduced or silenced in TRI6 disruption mutants [21]. Relative expression patterns of TRI13 and TRI101 were greatly down-regulated (p,0.001) after 3 and 5 days of confrontation with G. irregulare isolates, respectively ( Fig. 4B and D). TRI101 encodes a trichothecene 3-O-acetyltransferase that acetylates the C-3 of various Fusarium trichothecenes, converting them to less toxic products [22,23]. In biosynthesis of DAS, the TRI101 acetyltransferase catalyzes the conversion of isotrichodermol to isotrichodermin [22,23].
Control of mycotoxin-producing fungal pathogens has largely relied on the use of chemicals [9]. Modulation of trichothecene biosynthetic gene expression by AMF may be a safer way of limiting mycotoxin production. The impact of AMF on plant pathogenic fungi has been studied under ecological conditions and in a large number of host-pathogen interactions [12,24]. These interactions can be direct, such as competition with the pathogen, or indirect, including (1) alleviation of abiotic stress such as enhanced nutrition of the host plant, (2) biochemical induced changes, and (3) interactions with microbiota in the rhizosphere . Most of the direct effects have been a result of AMF interacting with pathogens in the rhizosphere in which complex associations exist among plant roots, soil, and microorganisms [11]. However, changes in plant root physiology due to AMF association are certain to have significant impacts on the rhizosphere microflora through alteration of root exudates and other nutrient-related mechanisms [12,24]. A direct interaction between G. irregulare and F. oxysporum has been studied with axenic system designed by St-Arnaud et al., (1995) [25] in which G. irregulare altered the growth of F. oxysporum. In this study, we also showed that G. irregulare had a significant inhibitory effect on growth of a virulent and mycotoxinproducing isolate of F. sambucinum. In addition, G. irregulare significantly down-regulated three trichothecene biosynthetic genes, TRI4, TRI13, TRI101. Confrontation with G. irregularre increased the expression of the trichodiene synthase gene TRI5, which may be a response to the down-regulation of the P450 oxygenase gene TRI4. Up-regulation of TRI6, a G. irregulare regulatory gene in trichothecene biosynthesis, may also contribute to the increase in TRI5 expression [26]. Our study confirms and demonstrates AMF influence on the growth of F. sambucinum and furthermore have an effect on mycotoxin biosynthetic gene expression. Further chemical analyses will be required to see if the changes observed in trichothecene gene expression in dual confrontation result in a change in trichothecene production. It has been reported that signaling volatile organic compounds known as semiochemicals have effects in plant protection [27]. For example, the production of these semiochemicals by plant tissue can be induced by fungal infection [28]. The production of some semiochemicals directly by some fungi and bacteria was also shown, as well as the possibility of transformed roots to produce semiochemicals [29]. Thus, the effect of G. irregulare on gene expression in F. Sambucinum could result from a direct effect of the AMF, but also from an indirect effect. G. irregulare could induce carrot roots to produce volatiles with activity against trichothecene production in F. Sambucinum.
We conclude that AMF can modulate mycotoxin gene expression of a plant fungal pathogen. This effect may be an important mechanism involved in biological control of plant pathogens. This finding brings a new level of understanding to plant-microbe interactions and control of plant pathogenesis.

Fungal strain and growth conditions
Fusarium sambucinum strain T5 was isolated and characterized from naturally infected potato plants (cultivar Riba) from a field located in the Montreal region in August 2008. The strain was grown and maintained on V-8 juice agar medium and in GYEP medium (2% glucose, 0.1% yeast extract, 0.1% peptone) [30]. Two isolates of the AMF G. irregulare DAOM-197198 and DAOM-234328 were grown in vitro in co-culture with Ri T-DNAtransformed carrot roots (Daucus carota L.) on a minimal (M) medium. Spores for both isolates of G. irregulare were collected from plates by dissolving the Gellan gum as described [31].

DNA extraction, PCR amplification and sequencing
DNA was extracted from freshly harvested fungal mycelium grown in liquid GYEP medium for 2-4 days. Fungal mycelium was dried and ground using mortar and pestle and DNA was extracted with DNeasy Plant Mini Kit (Qiagen, Canada) following the manufacturer's instructions. PCR amplifications of trichothecene genes: TRI1, TRI3, TRI4, TRI4, TRI6, and TRI101 were performed on G. pulicaris (anamorph: F. sambucinum) strain T5 using the primer sets listed (Table 1). PCR amplifications in 50 mL all contained: 16 Taq buffer, 0.25 mM of each dNTP, 0.5 mM of each primer, 1 U of Taq DNA polymerase (Fermentas), and 50 ng of DNA template. Reactions were carried out using a thermal cycler EP Mastercycler S (Eppendorf) under the following parameters: pre-denatured at 94uC for 1 min, followed by 30 cycles of denaturation at 94uC for 30 sec, annealing at 51uC for 30 sec and elongation at 72uC for 90 sec, and a final, elongation at 72uC for 5 min. For each trichothecene gene, a negative control without DNA template was performed. PCR amplificons were visualized on a 1% agarose gel stained with ethidium bromide and visualized under UV light. Reactions that showed clear amplification bands were sequenced at the Genome Quebec Innovation Center (Montreal, Qc), using the specific primers (Table 1). Two sequencing reactions were performed for each PCR amplicon. Recovered sequences were assembled and analyzed using Vector NTI software (Invitrogen) and compared to the NCBI database using Nucleotide BLAST search. Nucleotide sequences were deposited in the EMBL nucleotide sequence database under the accession numbers: HQ445900 for ITS and HQ445905 to HQ445907 for TRI genes.

Chemical analysis of the trichothecenes
To induce mycotoxin production in liquid culture, a two-stage medium protocol modified from the method of Miller and Blackwell (1986) was employed [32,33]. The cultures were grown at 25uC on a rotary shaker at 200 rpm in the dark. After 7 days of incubation in the second stage medium, a 5 ml aliquot containing fungal material was extracted with 2 ml ethyl acetate. Extracts were analyzed with gas chromatography and low resolution mass spectrometry (GCMS) using a Hewlett Packard 6890 gas chromatograph fitted with a HP-5MS column (30 m60.25 mm film thickness) and a 5973 mass detector. Trichothecenes were identified by comparison of retention time and mass spectra with standard compounds.

Dual culture assays
The confrontation cultures between G. irregulare and F. sambucinum were performed in vitro using two-compartment Petri dishes (100615 mm). One compartment was filled with 25 ml GYEP agar medium (2% glucose, 0.1% yeast extract, 0.1% peptone and 2% agar) for F. sambucinum. The other compartment of the plates was gently filled with 25 ml M medium. GYEP and M media were connected by a bridge over the separation of the two compartments that allowed fungal hyphae to cross from one compartment to the other. Approximately 2 cm 2 of G. irregulare (isolates DAOM-197198 and DAOM-234328) and transformed carrot root co-cultures were individually transferred into each compartment containing M medium. Because AMF grow slowly, plates were incubated at 25uC for 4 weeks until the AMF hyphae grew to reach the bridge. The cultures were examined weekly and carrot roots were trimmed aseptically to prevent their growth into the distal compartment. Controls consisted of Ri T-DNA transformed carrot without AMF and F. sambucinum alone (neither roots nor AMF). An agar disk of 0.5 cm diameter containing F. sambucinum strain T5 was used to inoculate the distal compartment containing GYEP agar medium on the side opposite G. irregulare. Additional controls were performed used a disk of the F. sambucinum adjacent to M medium alone (Fs+M) or with carrot-

RNA isolation and real-time qRT-PCR assays
Total RNA was extracted from the mycelia of three biological replicates of F. sambucinum. Fungal material was prepared as described in the DNA extraction section. Total RNA was isolated using TRIZOL Reagent according to the manufacture's instruc-tions. cDNA libraries were constructed by RevertAid 270 H Minus M-MuLV kit (Fermentas) according to the manufacture's instructions. Real-time (RT) PCR reactions were performed in a volume of 10 ml containing 2 ml water, 1 ml of each primer, 1 ml cDNA and 5 ml CYBR green Maxima SYBR Green/ROX qPCR Master Mix. A liquid handling Workstation epMotion 5070 (Eppendorf) was used to optimize RT-PCR assays with small reaction volumes. All genes were run in triplicate on each plate and 3 biological replicates of each treatment were performed. A negative control using Mili-Q water was prepared for each sample. RT-PCR was run on EP RealPlex MasterCycler (Eppendorf) using the following conditions: an initial denaturation step at 95uC for 10 min followed by 39 cycles of 95uC for 15 sec, 60uC for 45 sec (annealing and extension). A final extension was carried out by 95uC for 15 sec followed by 60uC for 1 min. A melting curve was performed from 55 to 95uC with a 0.2uC reading interval. We used two house-keeping genes, ß-tubulin elongation factor EF1a, in our RT-PCR assays. All samples had only a single peak, indicating a pure RT-PCR product and no contamination or primer dimer formation. Data analysis was performed on REST 2009 Software available at http://www.gene-quantification.de/ rest-2009.htm as described below in in statistical analysis section.

Statistical analysis
Analysis of variance was used to examine the significant difference on effects of the AMF G. irregulare isolate DAOM-197198 on growth of F. sambucinum and Shannon diversity indices between different treatments, and post-hoc comparisons between the treatments were done by the Tukey test using SPSS software v. 17 (SPSS Inc., Chicago, Illinois).
We used Relative Expression Software Tool (REST) for groupwise comparison and statistical analysis of relative expression results as described in Pfaffl et al., [34]. The relative expression ratio of a target gene is computed, based on its real-time PCR efficiencies and the crossing point (CP) difference of an unknown sample versus a control. The purpose of this test is to determine whether there is a significant difference between samples and controls, while taking into account issues of reaction efficiency and reference gene normalization. We used the hypothesis test P(H 1 ) that represents the probability of the alternate hypothesis that the difference between sample and control groups is due only to chance. The hypothesis test performs a least 2000 times of random reallocations of samples and controls between the groups. Statistical difference are significant when p,0.05.